Possible to use regex to move age-level but not genre to a custom column?

[iamagloworm]

Hello, I have populated my tags column with age-levels like 04-07, 05-10, 16-20 etc. along with genres like fairytale, picture book, animals

I would like to move the age-level to a new column. In the bulk search and replace, I can use a regex like

(\b\d{2}-\d{2}\b)

to identify the string I want, and in the replace I try something like

\1

in the new field.

but it copies the genre tags too...

Any advice?

Also, I would prefer to move them, instead of copy, if that is possible.

Thanks!

[kjdavies]

my calibre is tied up right now or I'd actually try this, but... a regex like this might work:

find the books to update:

tags:"~\b\d\d-\d\d\b"

search field: tags

find:

^(.*?)((\b\d\d-\d\d\b))+(.*?)$

replace:

\2,

into:

#myagetags

The comma will force an empty tag at the end, but that should fall off (empty tags disappear automatically).

This should be nondestructive to the original tags. After your satisfied it works,

search field: tags
find:

\b\d\d-\d\d\b

replace:

(i.e. replace with nothing)

should strip the age tags. Or you could go into the tag manager and manually select them, then remove from there.

[iamagloworm]

Thanks for your help. That is how I would expect it to work, but it doesn't. I am wondering if it is something to do with calibre's separator characters

:::

?
See the attached screenshot if. Am I missing something?

[chaley]

The problem is that for multiple valued columns like tags, search & replace iterates over the individual values (each tag). It doesn't operate on all of them at once. Thus the regular expression is applied to each tag, not the string of all tags.

You can get around this by using *template as the input field, where the template is something like {tags}. The result of a template is a single value, not multiple values, even if it contains commas. This causes S&R to operate on the single value, not individual ones.

You could even get fancier and have the template generate the value you actually want in the destination column, eliminating the regex. This would be useful mostly if splitting apart the tags requires some logic & conditionals, which doesn't seem to be the case here.

[iamagloworm]

Originally Posted by chaley

You can get around this by using *template as the input field, where the template is something like {tags}. The result of a template is a single value, not multiple values, even if it contains commas. This causes S&R to operate on the single value, not individual ones.

This works pretty well, however, if there is no age-level tag is copies the genre tags. Is there a way to prevent this?

My cave-man workaround is just to exclude any books missing age-level tags...

Attached is my config that mostly works:

[chaley]

Originally Posted by iamagloworm

This works pretty well, however, if there is no age-level tag is copies the genre tags. Is there a way to prevent this?

My cave-man workaround is just to exclude any books missing age-level tags...

Attached is my config that mostly works:


Attachment 203771

To avoid the pre-selection you need to "get fancy".

This works for me. It depends on the tag always being nn-nn with no other leading or trailing characters. It uses the first one it finds. If it desn't find any then it returns ''

The General Program Mode template is

program:
	for t in $tags:
		if "^\d\d-\d\d$" in t then
			return t
		fi
	rof;
	return ''

The S&R is

[iamagloworm]

Awesome, thank you so much!

[kjdavies]

Originally Posted by chaley

The problem is that for multiple valued columns like tags, search & replace iterates over the individual values (each tag). It doesn't operate on all of them at once. Thus the regular expression is applied to each tag, not the string of all tags.

You can get around this by using *template as the input field, where the template is something like {tags}. The result of a template is a single value, not multiple values, even if it contains commas. This causes S&R to operate on the single value, not individual ones.

You could even get fancier and have the template generate the value you actually want in the destination column, eliminating the regex. This would be useful mostly if splitting apart the tags requires some logic & conditionals, which doesn't seem to be the case here.

That's what I get for not trying it (calibre was busy doing something).

If examining tags works on a per-tag basis...

Oh. If the tag doesn't match the pattern, it'll still output the original value. You'll copy all the tags.

On the other hand, and this would probably be pretty wasteful, wouldn't this work?

find: ^(\d\d-\d\d)$
replace: \1
target: #agetag
(concatenate at tags -- ::: inserted)

Then to remove the other tags, I'd probably go to the tag manager, but I suspect the regex replace below could work (probably slowly)

find: ^(\d\d-\d\d)$
replace:
target: tags (replace with nothing)

If I'm not mistaken -- and I totally could be -- the empty tags should drop out and leave you with just the not-empty tags.

(calibre is busy moving a bunch of files right now, or I'd actually try it and find out)